Matt Garman became CEO of Amazon Web Services on June 3, 2024, succeeding Adam Selipsky in what Amazon described as a planned succession. The change did not signal that AWS was failing. It reflected a leadership handoff at a strategically difficult moment: AWS remained the largest and most profitable major cloud platform, but Microsoft Azure and Google Cloud had stronger perceived momentum in generative artificial intelligence.
Garman inherited a dominant cloud business that needed to turn its infrastructure advantage into an AI advantage. His response has been to make AWS a full-stack AI provider—combining custom chips, Bedrock, model partnerships, agents, and large capacity investments—while preserving AWS’s traditional strengths in enterprise breadth, reliability, security, and cost control.
As an Amazon Associate I earn from qualifying purchases.
A planned succession at a strategic inflection point
Amazon announced that Garman would take over AWS from Selipsky effective June 3, 2024. Amazon CEO Andy Jassy said Selipsky had been expected to lead AWS for several years while helping develop the next generation of leadership.
That explanation matters. There is no supported evidence that Selipsky was fired because of poor performance. His approximately three-year tenure included the post-pandemic normalization of cloud spending, customer cost-cutting, and the sudden rise of generative AI. The leadership change occurred during a major strategic transition, but the available evidence supports a planned succession rather than a confirmed turnaround appointment.
#1 Best Overall
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
At the same time, market observers described AWS as being at a crossroads over AI. Microsoft had associated Azure closely with OpenAI and ChatGPT, while Google brought deep AI research and its own model capabilities. AWS had important machine-learning infrastructure and services, but its public AI narrative appeared less dominant.
Those two facts can both be true: AWS planned its succession, and the incoming CEO faced a serious strategic challenge.
Amazon’s leadership announcement said AWS had reached a $100 billion annual revenue run rate around the transition.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallWho is Matt Garman?
Garman was an unusually logical internal successor for the problem AWS faced. He joined Amazon in 2006 as an MBA intern and held roles across product management, engineering, operations, compute, storage, sales, marketing, and global services.
His career was built around cloud infrastructure and enterprise adoption—not primarily around consumer software or frontier AI research. That distinction is important. AWS’s challenge was not simply to create the most popular chatbot. It had to:
- Supply scarce computing capacity and advanced networking;
- Design and deploy custom AI chips;
- Offer multiple foundation models to enterprise developers;
- Support training and inference workloads;
- Integrate AI with databases, storage, security, analytics, and applications; and
- Convince customers to move AI experiments into production.
Garman’s experience across both the technical stack and enterprise sales side made him suited to that operating challenge. His official biography details his progression through AWS leadership.
What “AWS at a crossroads” really meant
The phrase was broader than “AWS was behind in AI.” It described several related decisions.
AI mindshare and model leadership
Microsoft had an early narrative advantage through its relationship with OpenAI. Google had major research capabilities and proprietary models. AWS had to avoid confusing cloud-market leadership with model-market leadership.
These are separate contests:
- Model leadership: developing frontier foundation models;
- Infrastructure leadership: supplying chips, servers, networking, and data centers;
- Platform leadership: helping businesses build, govern, and operate AI applications; and
- Cloud leadership: providing the broadest general-purpose enterprise platform.
AWS could lag in public model mindshare while remaining powerful in infrastructure and enterprise distribution. Conversely, it could win infrastructure consumption without owning the most valuable AI applications.
Rank #2
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
The economics of AI infrastructure
AI workloads require expensive accelerators, networking, power, cooling, and data-center capacity. That creates a difficult investment decision. If AWS builds too little capacity, customers may go elsewhere. If it builds too much, Amazon could be left with underutilized infrastructure.
There is also a chip trade-off. Relying heavily on Nvidia gives AWS access to a mature ecosystem but leaves it dependent on a powerful supplier. Developing custom silicon may improve price-performance and supply control, but it creates software, compatibility, and execution risks.
Free tools Windows power users keep installed
One-click scans. No signup required.
Growth versus margins
AI can increase cloud consumption, but it can also be expensive to provide. AWS must determine whether customers will pay enough for training, inference, and managed AI services to justify the required capital investment.
The important questions are whether AI creates durable new workloads, how much demand comes from a small number of laboratories, whether custom chips improve margins, and whether model providers capture most of the value while cloud companies compete mainly on price and capacity.
AWS’s strengths when Garman took over
AWS did not begin the Garman era from a position of weakness. It had a large enterprise customer base, a mature global footprint, an extensive partner ecosystem, a broad managed-services portfolio, and years of experience operating at hyperscale.
It also had established relationships with enterprise buyers, developers, and regulated industries. Those relationships are valuable in AI because production deployments require far more than model access. They require identity, networking, data storage, security, monitoring, compliance, and reliable operations.
AWS could offer AI through several routes:
- Nvidia-based EC2 instances;
- Amazon-designed Trainium and Inferentia chips;
- Amazon Bedrock for managed foundation models;
- Amazon’s own model families;
- Third-party models from companies including Anthropic, Google, and OpenAI; and
- Managed services for agents and application development.
Garman’s AI strategy
1. Make AWS useful regardless of the winning model
Bedrock is central to AWS’s model-neutral strategy. Rather than forcing customers to use one Amazon model, AWS offers access to multiple foundation models through a managed service.
This gives customers choice and allows AWS to benefit when customers prefer another model provider. It also encourages companies to keep data, governance, and application workloads inside AWS while they test or switch models.
The trade-off is differentiation. If AWS becomes mainly an interchangeable access layer, model companies may retain more of the value. Customers may also find it easier to replace one cloud abstraction layer with another.
Rank #3
- Your favorite music and content – Play music, audiobooks, and podcasts from Amazon Music, Apple Music, Spotify and others or via Bluetooth throughout your home.
- Alexa is happy to help – Ask Alexa for weather updates and to set hands-free timers, get answers to your questions and even hear jokes. Need a few extra minutes in the morning? Just tap your Echo Dot to snooze your alarm.
- Keep your home comfortable – Control compatible smart home devices with your voice and routines triggered by built-in motion or indoor temperature sensors. Create routines to automatically turn on lights when you walk into a room, or start a fan if the inside temperature goes above your comfort zone.
- Do more with device pairing – Fill your home with music using compatible Echo devices in different rooms, or create a home theatre system with Fire TV.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
Amazon said in its Q4 2025 results that Bedrock offered more than 20 fully managed models and supported experimentation without requiring customers to rewrite applications.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →2. Build custom chips
Trainium and Graviton are key parts of AWS’s attempt to control more of its cost structure. Amazon said Trainium and Graviton together exceeded a $10 billion annual revenue run rate in its Q4 2025 materials. It also said Trainium2 powered Project Rainier, a large AI cluster used by Anthropic, and that Trainium3 demand was strong.
The strategic logic is straightforward: custom chips could lower cost per token, reduce dependence on Nvidia, and let AWS optimize hardware and software together.
But custom silicon is not automatically a Nvidia replacement. Its value depends on software compatibility, developer tools, workload performance, supply, networking, and availability. A cheaper chip that developers struggle to use—or cannot obtain at scale—is not a complete alternative.
3. Move customers from pilots to production
The larger opportunity is not simply letting companies demonstrate chatbots. AWS wants businesses to run recurring workloads such as customer service, software development, document processing, fraud detection, supply-chain operations, internal search, and automated agents.
Recommended Free Tools
AWS’s re:Invent 2025 announcements emphasized agents, Bedrock AgentCore, model families, Trainium3, and AI Factories. That points to a strategy extending from chips and compute to deployment, governance, and ongoing operations.
4. Use partnerships to fill gaps and anchor demand
AWS has combined internal development with relationships involving Anthropic, Oracle, Intel, and OpenAI. These partnerships can generate cloud consumption, improve technical credibility, and give AWS access to important enterprise or AI workloads.
They also create risks. Large AI customers can have significant bargaining power, consume substantial capacity, and remain competitors as well as customers. A partnership can increase revenue without proving that the infrastructure produces attractive returns.
What the evidence shows so far
Amazon’s subsequent disclosures suggest that AWS responded aggressively rather than remaining stuck at the crossroads.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Rank #4
- Meet Echo Dot Max: Experience rich room-filling sound that automatically adapts to your space and fine-tunes playback. Features a built-in smart home hub and Omnisense technology for highly personalized experiences.
- Music to your ears: With nearly 3x the bass versus Echo Dot (2022 release), it fits beautifully in any space, delivering your personal sound stage with deep bass and enhanced clarity. Listen to streaming services, such as Amazon Music, Apple Music, Spotify, and SiriusXM. Encore!
- Do more with device pairing: Connect compatible Echo smart speakers and smart displays in different rooms, or pair with a second Echo Dot Max to enjoy even richer sound
- Simple smart home control: Set routines, pair and control lights, locks, and thousands of smart home devices that work with Alexa without needing a separate smart home hub. With Omnisense technology, you can activate routines via temperature or presence detection.
- Say goodbye to drop-offs and buffering - With eero Built-in, Echo Dot Max doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
- Amazon said AWS AI revenue exceeded a $15 billion annual run rate in the first quarter of 2026.
- Amazon reported AWS operating income of $12.5 billion in the fourth quarter of 2025, compared with $10.6 billion in the fourth quarter of 2024.
- Amazon reported AWS operating income of $16.6 billion in the second quarter of 2026, compared with $10.2 billion in the second quarter of 2025.
- Amazon said Bedrock usage was growing and that customers were increasingly using it for production workloads.
- Amazon said Trainium and Graviton had exceeded a combined $10 billion annual revenue run rate.
These are encouraging indicators, but they require careful interpretation. The AI revenue figures are company-defined run-rate measures, not a separately audited reporting segment. They do not by themselves reveal margins, customer concentration, discounts, utilization, or return on invested capital.
Microsoft reported 39% growth in Azure and other cloud services in fiscal Q2 2026. Comparisons must be made carefully because the companies use different fiscal calendars, reporting definitions, and business mixes. Still, the figure illustrates that AWS continues to face a strong rival with significant AI momentum.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to judge whether Garman’s strategy is working
AWS should not be judged by a single headline number. Investors, technology executives, and cloud customers should watch six areas.
AWS growth quality
Is growth coming from broad enterprise consumption and production AI, or from a few large commitments and temporary capacity shortages? Revenue growth is more durable when it is diversified across customers and workloads.
AI revenue quality
Managed services and recurring inference may be economically different from low-margin infrastructure resale, one-time training projects, internal Amazon usage, or future capacity commitments. Amazon’s run-rate figures are useful signals, but not complete profitability measures.
Custom-chip adoption
The relevant test is not whether Trainium generates revenue. It is whether customers choose it for attractive workloads, whether the software stack supports mainstream frameworks, whether capacity is available, and whether the chips improve AWS’s economics.
Production adoption
The strongest evidence would be sustained inference and application usage after pilots, credits, and promotional periods end. Conference demonstrations and announced partnerships are weaker evidence than repeat consumption tied to measurable business outcomes.
Capital efficiency
AI infrastructure requires heavy spending. Over time, AWS must show durable operating-income growth, improving unit economics, sufficient utilization, manageable power and data-center costs, and acceptable returns on the capital invested.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsEcosystem resilience
A strong AWS position would allow customers to use multiple models, Nvidia and AWS chips, open-source and proprietary systems, containers, serverless tools, and hybrid architectures. Customer choice can increase AWS’s usefulness, although it can also make the platform easier to substitute.
Best Value
- Meet Echo Studio: Redesigned in a compact size that is 40% smaller than the original to deliver immersive spatial sound. Dolby Atmos adds space, clarity, and depth to make your entertainment even better. Enjoy powerful bass and crystal-clear vocals. Features a built-in smart home hub and Omnisense technology for highly personalized experiences.
- It's like a concert in your living room: Experience immersive sound thanks to spatial audio and Dolby Atmos in a compact design that fits easily into any living space. With room adaptation technology, Echo Studio analyzes the acoustics of your room, fine-tuning playback for optimal sound no matter where it's placed.
- Do more with device pairing: Play music across multiple Echo devices for multi-room music, or pair with a second Echo Studio for even more powerful sound.
- Simple smart home control: Set routines, pair and control lights, locks, and thousands of devices that work with Alexa without needing a separate smart home hub. With Omnisense technology, you can activate routines via temperature or presence detection.
- Say goodbye to drop-offs and buffering: With eero Built-in, Echo Studio doubles as a mesh wifi extender, adding up to 1,000 sq. ft. of wifi coverage to your existing eero network.
The unresolved risks
Model neutrality may limit differentiation. Bedrock can monetize demand for competing models, but AWS may become the infrastructure utility beneath more valuable model companies.
Custom silicon may not displace Nvidia. AWS uses both Nvidia hardware and its own chips. Trainium’s success depends on software, supply, and performance—not just announced specifications.
AI demand may be economically immature. Customers can genuinely need more compute while still failing to earn enough from their AI applications to support long-term spending.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Large customers may create concentration risk. Anthropic, OpenAI, and other AI developers can provide major demand, but they also possess negotiating power and may eventually compete with AWS at other layers.
AWS could win infrastructure while losing the application layer. Customers might use AWS for compute but build their applications around another company’s models, enterprise software, data tools, or developer ecosystem.
Capital spending could pressure returns. Capacity investment can protect AWS from shortages, but slower adoption or falling prices could leave Amazon with expensive infrastructure and weaker returns.
Bottom line
Matt Garman took over AWS because Amazon’s planned succession coincided with a major change in the cloud industry. AWS was not a broken business: it was already a $100 billion-scale, highly profitable platform with substantial enterprise advantages.
Recommended Free Tools
But generative AI challenged AWS’s public momentum and economics. Garman’s answer has been to use AWS’s existing strengths—scale, infrastructure, enterprise relationships, and operational expertise—to build a broader AI stack involving Bedrock, custom chips, external models, agents, and major capacity commitments.
The early evidence suggests AWS has become more aggressive and more relevant in AI. The definitive test is still financial and operational: whether AI demand becomes diversified, durable, and profitable rather than merely requiring Amazon to spend heavily to remain competitive.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
